23440034522

Committed 23 Mar 2026 01:35PM UTC coverage: 99.266% (-0.4%) from 99.66%

Build # 23440034522

Build Type

Pull #488

github

Committed by

web-flow

Commit Message

Merge 7244aa3c4 into 1b53720fc

Pull Request Pull Request #488: Add multilinear regression

Coverage Stats

142 of 171 new or added lines in 14 files covered. (83.04%)

7 existing lines in 3 files now uncovered.

8657 of 8721 relevant lines covered (99.27%)

222.35 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

87.5

/src/Statistics/Regression/PowerLaw.php

<?php

namespace MathPHP\Statistics\Regression;

use MathPHP\Exception;

/**
 * Power law regression (power curve) - Least squares fitting
 * http://mathworld.wolfram.com/LeastSquaresFittingPowerLaw.html
 *
 * A functional relationship between two quantities,
 * where a relative change in one quantity results in a proportional
 * relative change in the other quantity,
 * independent of the initial size of those quantities: one quantity
 * varies as a power of another.
 * https://en.wikipedia.org/wiki/Power_law
 *
 * y = Axᴮ
 *
 * Using least squares fitting: y = axᵇ
 *
 *     n∑⟮ln xᵢ ln yᵢ⟯ − ∑⟮ln xᵢ⟯ ∑⟮ln yᵢ⟯
 * b = --------------------------------
 *           n∑⟮ln xᵢ⟯² − ⟮∑⟮ln xᵢ⟯⟯²
 *         _                    _
 *        |  ∑⟮ln yᵢ⟯ − b∑⟮ln xᵢ⟯  |
 * a = exp|  ------------------  |
 *        |_          n         _|
 *
 * @phpstan-import-type SimpleLinearResultModel from Methods\LeastSquares
 * @phpstan-import-type PolynomialResultModel from Methods\LeastSquares
 */
class PowerLaw extends ParametricRegression
{
    use Models\PowerModel;
    /** @use Methods\LeastSquares<SimpleLinearResultModel> */
    use Methods\LeastSquares;

    /** @var float */
    protected $a;

    /** @var float */
    protected $b;

    /**
     * @param list<float> $array
     * @return SimpleLinearResultModel
     */
    protected function createResultModel(array $array): array
    {
        return $this->createSimpleLinearResultModel($array);
    }

    /**
     * Calculate the regression parameters by least squares on linearized data
     * ln(y) = ln(A) + B*ln(x)
     *
     * @throws Exception\BadDataException
     * @throws Exception\IncorrectTypeException
     * @throws Exception\MatrixException
     * @throws Exception\MathException
     */
    public function calculate(): void
    {
        // Linearize the relationship by taking the log of both sides.
        $x’ = \array_map('\log', $this->xs);
        $y’ = \array_map('\log', $this->ys);

        // Perform Least Squares Fit
        $linearized_parameters = $this->leastSquares($y’, $x’)->getColumn(0);

        // Translate the linearized parameters back.
        $this->a = \exp($linearized_parameters[0]);
        $this->b = $linearized_parameters[1];

        $this->parameters = [$this->a, $this->b];
    }

    /**
     * Evaluate the regression equation at x
     * Uses the instance model's evaluateModel method.
     *
     * @param  float $x
     *
     * @return float
     */
    public function evaluate(float $x): float
    {
        return $this->evaluateModel($x, $this->parameters);
    }
}

1	<?php
2
3	namespace MathPHP\Statistics\Regression;
4
5	use MathPHP\Exception;
6
7	/**
8	* Power law regression (power curve) - Least squares fitting
9	* http://mathworld.wolfram.com/LeastSquaresFittingPowerLaw.html
10	*
11	* A functional relationship between two quantities,
12	* where a relative change in one quantity results in a proportional
13	* relative change in the other quantity,
14	* independent of the initial size of those quantities: one quantity
15	* varies as a power of another.
16	* https://en.wikipedia.org/wiki/Power_law
17	*
18	* y = Axᴮ
19	*
20	* Using least squares fitting: y = axᵇ
21	*
22	* n∑⟮ln xᵢ ln yᵢ⟯ − ∑⟮ln xᵢ⟯ ∑⟮ln yᵢ⟯
23	* b = --------------------------------
24	* n∑⟮ln xᵢ⟯² − ⟮∑⟮ln xᵢ⟯⟯²
25	* _ _
26	* \| ∑⟮ln yᵢ⟯ − b∑⟮ln xᵢ⟯ \|
27	* a = exp\| ------------------ \|
28	* \|_ n _\|
29	*
30	* @phpstan-import-type SimpleLinearResultModel from Methods\LeastSquares
31	* @phpstan-import-type PolynomialResultModel from Methods\LeastSquares
32	*/
33	class PowerLaw extends ParametricRegression
34	{
35	use Models\PowerModel;
36	/** @use Methods\LeastSquares<SimpleLinearResultModel> */
37	use Methods\LeastSquares;
38
39	/** @var float */
40	protected $a;
41
42	/** @var float */
43	protected $b;
44
45	/**
46	* @param list<float> $array
47	* @return SimpleLinearResultModel
48	*/
49	protected function createResultModel(array $array): array
50	{
NEW 51	return $this->createSimpleLinearResultModel($array);	×
52	}
53
54	/**
55	* Calculate the regression parameters by least squares on linearized data
56	* ln(y) = ln(A) + B*ln(x)
57	*
58	* @throws Exception\BadDataException
59	* @throws Exception\IncorrectTypeException
60	* @throws Exception\MatrixException
61	* @throws Exception\MathException
62	*/
63	public function calculate(): void
64	{
65	// Linearize the relationship by taking the log of both sides.
66	$x’ = \array_map('\log', $this->xs);	8✔
67	$y’ = \array_map('\log', $this->ys);	8✔
68
69	// Perform Least Squares Fit
70	$linearized_parameters = $this->leastSquares($y’, $x’)->getColumn(0);	8✔
71
72	// Translate the linearized parameters back.
73	$this->a = \exp($linearized_parameters[0]);	8✔
74	$this->b = $linearized_parameters[1];	8✔
75
76	$this->parameters = [$this->a, $this->b];	8✔
77	}
78
79	/**
80	* Evaluate the regression equation at x
81	* Uses the instance model's evaluateModel method.
82	*
83	* @param float $x
84	*
85	* @return float
86	*/
87	public function evaluate(float $x): float
88	{
89	return $this->evaluateModel($x, $this->parameters);	5✔
90	}
91	}

markrogoyski / math-php / 23440034522

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous